Ahrefs’ 2026 roundup of the researchers producing the most original AI-search studies is more than a who’s-who: it reinforces a clear operational takeaway for marketers. The research shows AI Overviews are already siphoning meaningful click volume, LLM citation behavior is unstable and often decoupled from classic ranking signals, and simple technical controls such as schema or /llms.txt rarely move the needle on their own. If you rely on organic search traffic, these are not theoretical risks — they change how you should measure, test and prioritise investment.

How Ahrefs built the list — and why the methodology matters

Ahrefs started with a longlist of 87 names and counted only studies that met four tests: original data collected or analyzed by the author, a focus on AI search (how assistants cite and recommend content), a meaningful sample size (for example ~50+ prompts or multi-site analytics), and a published write-up between 2023 and 2026. Rankings were by the number of qualifying studies, with backlink counts used to break ties.

That method privileges empirical, repeatable work: large samples, experiments and measurable outcomes. The trade-off is a bias toward volume — teams that publish many data-heavy studies (Ahrefs, SISTRIX, DEJAN, Peec AI and similar) rise to the top, while occasional but influential theoretical or product-level findings that don’t fit the ‘study’ definition may be underrepresented.

What the top researchers are consistently testing

The list highlights a handful of recurring, evidence-backed themes that should shape marketing decisions today.

  • AI Overviews and click loss. Several analyses cited in the roundup estimate material click reductions when AI Overviews appear. One Ahrefs study referenced a roughly 34.5% decline in clicks across hundreds of thousands of keywords.
  • Citation mechanics are volatile. Large datasets show LLM citations shift often across engines (ChatGPT, Perplexity, Google AI), and cited sources frequently don’t align with organic ranking order.
  • Small technical signals under-deliver. Multiple tests reported in the compilation found limited measurable impact from schema changes and /llms.txt in many real-world samples; in some large samples most /llms.txt files weren’t read at scale.
  • Brand and freshness matter more. Researchers repeatedly flag brand authority and fresh content as stronger predictors of being cited by AI assistants than isolated technical tweaks.

Who’s publishing the most—and what that tells you

The top contributors are heavy on organizations that publish frequent, data-heavy research. Kevin Indig led the list with 36 qualifying studies; Ahrefs’ Xibeijia Guan and Louise Linehan together account for dozens of large-sample reports. Agencies and analytics firms such as DEJAN, SISTRIX, Profound and Peec AI are also prominent for their citation and crawling experiments. For practitioners, the value of these authors is practical: they publish reproducible methods and large samples you can benchmark against.

Immediate implications — what marketing and SEO teams should do

  • Instrument AI-driven referrals. Add explicit tracking for references from AI assistants and Overviews (separate event tags or referral classifications). Traditional organic metrics will miss much of the shift.
  • Treat AEO as a separate performance channel. Visibility in organic SERPs helps but doesn’t guarantee citation by LLMs. Build distinct measurement and success criteria for AEO/GEO outcomes.
  • Prioritise brand and topical authority. Invest in authoritative, timely content and brand signals that LLMs are more likely to surface than isolated micro-optimisations.
  • Use schema and /llms.txt, but don’t over-allocate budget. These controls should be part of hygiene work, not the core of your AEO strategy, given limited measurable impact in many tests.
  • Expect and plan for volatility. Citation sets change across engines and over time. Maintain diversified discovery: organic SEO, branded channels, and content that clearly cites sources, data and unique analysis.
  • Make reproducible testing mandatory. Replicate small experiments with clear sample sizes and methods that mirror your vertical before scaling tactics across your content program.

Following the researchers highlighted by Ahrefs is useful because many publish replicable, large-sample studies. But remember: the list rewards frequency. Judge each study on its sample relevance and method quality before applying its conclusions to your business.

What to watch next — and the practical next steps

Key signals to monitor: whether crawler signals (including standardized /llms.txt or equivalent) gain adoption and demonstrable effect in broader samples; continued measurement of AI Overviews’ impact on clicks; and whether platforms provide clearer attribution for AI-driven referrals. For teams running content operations, start by instrumenting AI referrals, designing two to three lightweight replication tests relevant to your vertical, and reallocating a small portion of your SEO budget toward brand and topical authority measures rather than technical one-off fixes.

This is not an academic exercise: if AI Overviews keep suppressing clicks, you’ll need measurement that shows it, content that still compels conversions when cited, and distribution that doesn’t rely on a single discovery mechanism.